Image Processing of Pedestrian Behavior With Multiple Color Markers and Its Application to Crowd Evacuation Experiment

Abstract

This paper proposes a novel image processing method to extract pedestrian movements in a crowd from video images of the crowd seen from above. A large scale crowd evacuation experiment with 96 pedestrians was analyzed with this method. This method detects color markers on a cap of each pedestrian and makes trajectories of pedestrians by clustering neighboring markers and tracking them during the sequential video frames. The color combination of clustered markers indicates the body height of each pedestrian to result in a more accurate coordinate transform. This method successfully tracked the pedestrians to make their trajectories with the positioning error under 40 mm. Over 99 percent of pedestrian's position data were automatically calculated without manual operation thanks to the redundancy of multiple markers and the marker clustering. In the merging experiment, the flow rate through an opening decreased due to turning off after the opening and merging with pedestrians in the path. Detailed pedestrian behavior revealed that walking direction and walking speed were strongly linked to the interaction between neighboring pedestrians such as the distance and the relative velocity between them.

Proposal of a classification method for public space usage based on pedestrian trajectory data

“…Studies classified under Category (1) focus on topics such as behaviors, for example avoidance and interaction of individual pedestrians in various scales from laboratories/living rooms to urban spaces, [3][4][5][6][7] spatial perception and behavioral characteristics of visually impaired people, 8,9 and behavioral characteristics of pedestrian groups 10 and crowds. [11][12][13][14] These studies mainly aim to apply actual measurement results to the estimation of behavioral models, [15][16][17][18][19] improvement of pedestrian simulations, 15,20,21 marketing, 22 security, 23 and mobility support for disabled individuals. [24][25][26][27][28] Research classified under Category (2), on an architectural scale involves those on fire prevention equipment, 29 design guidelines for corridor spaces, [30][31][32] and exhibitions and mobility in museums.…”